Personalized Trusted Service Recommendation Method

The booming Web technology brings both convenience and information overload. To manually search for information through searching engines is not so intelligent as the information is too large now. The academic puts forward the personalized service recommendation technology, which can positively pick information from large amounts of information by study through people's hobbies and behavior. However, traditional technologies did not embrace this concept and often failed to recommend good information.A personalized recommendation method based on social network is proposed to improve the reliability and accuracy of service recommendation. First, it calculates the direct trust and indirect trust between friends using the trust relationship of the friends on social network,which lies a solid foundation for a trustworthy recommendation from a friend. Meanwhile, a preference trust is also proposed to meet the need for reliability and user's preference. Second,a cluster of similar services which have been called is made into a set of similar services.Thus increases the accuracy of recommendation. Finally, the indirect trust degree is predicted through direct trust degree of services already called. And the indirect trust degree is used in the future recommendations.MovieLens data set, a recommendation test set, is used in the simulation experiment which has accuracy rate, recall rate and F1_Measure as the evaluation index and the number of recommended user sets and the recommended list lengths as its variants. The result is compared with those of traditional methods and trust network methods.